Peter Fast

Selected case study

FAST Strength Software

Connect a training plan, completed workouts, history, and video review in one workflow.

In development · tracking remains experimental

Updated October 2026

Python · Django · JavaScript · PostgreSQL · Automated tests

The problem

Bring the plan, the completed workout, and video feedback into one workflow while keeping each lifter’s records under their control.

My contribution

  • Defined workout, history, and video-review workflows from coaching experience.
  • Developed the Django application with AI assistance and iterative source, test, and interface review.

Evidence you can inspect

The screen below comes from the implemented Django application’s October 21 local demonstration. It separates the planned workout from the sets actually saved.

In this fictional session, the plan is 3 × 5 at 40 kg. The completed entries are 5 × 40 kg, 6 × 45 kg, and 6 × 45 kg: 17 repetitions and 740 kg-reps. The last set was intentionally repeated in the demonstration.

The finished state, editable records, and explicit “Resume workout” action are visible. The screenshot contains no lift video and makes no claim about tracking accuracy.

Actual local application screen for Fictional Client A: a finished workout shows three logged sets, 17 repetitions, and 740 kilogram-repetitions.
Actual local capstone screen, October 21, 2026. All client and exercise data are fictional. Software remains in development. Open full evidence image

How I approached the work

  1. Define the workout and review workflows from training experience
  2. Build the data models, browser interactions, and access checks
  3. Use local demonstrations to review entry, correction, and recovery behavior
  4. Check the implementation with Windows and PostgreSQL test runs

Demonstrated outcomes

  • Implemented planned and one-off workout recording, training history, and video review with manual corrections.
  • Built owner-scoped records and explicit coaching access.

Limitations and unfinished work

  • The software is in development and is not a public client service.
  • Automated tracking is experimental; independent human labels and motion-accuracy validation are incomplete.
  • The screenshot documents a local fictional demonstration, not athlete results, hosted availability, or acceptance on physical phones. The production repository is private.

Explore further